[00:27] flacoste: /me lunches [00:27] oops [00:27] * mwhudson lunches [01:37] Anyone else seeing a spurious ordering-based failure in lib/lp/bugs/tests/../stories/bugs/xx-front-page-bug-lists.txt ? [02:22] can anyone give me a link where I can find more info. on writing tests for a change I just made, thankw [02:22] *thanks [04:59] Series/milestones/releases could really do with being unbroken. [05:00] wgrant: i think sinzui knows that [05:00] and i'm not sure anyone else understands things well enough to unfix them [05:01] you could argue any random change would likely be an improvement, but that's probably not actually true [05:01] I don't get why no time has been dedicated to making those important, obvious features less restrictive and broken. [05:02] * mwhudson eods [05:42] spm: Oh, crap, 499421 has happened again? [05:42] wgrant: yeah :-( [05:42] I was like WTF is this alert about!?!? [05:42] fortunately tom belives in docco in our alert scripts so enough clues to find the bug/solution [05:43] no idea if helpful but was the "PPA for Ubuntu Mozilla Daily Build Team" [05:44] At least the other one (499095) seems to be properly fixed now. [07:49] night [07:51] night poolie [08:26] good morning [08:57] Morgen [10:12] noodles775, james_w: hi [10:12] hi jml [10:13] james_w: may I join your call? [10:13] jml: fine by me [10:13] not sure how to do that though [10:14] james_w, right-click on me in your skype contacts and select "Invite to conference" [10:36] james_w, noodles775: I've started a gobby document on gobby.ubuntu.com, 'lp-daily-builds' === matsubara-afk is now known as matsubara [11:15] hi all [11:16] jml: did that second branch land yesterday? [11:17] I think that's the third time it's been put in ec2, perhaps it breaks tests? [11:30] gmb: Hi, and thanks for the reviewing. Could you go ahead and land lp:~maxb/launchpad/use-hashlib for me? I've set a commit message on the MP. My other branch (lp:~maxb/launchpad/stop-using-deprecated-sets) would merge-conflict with that so if you leave that one alone, I'll merge and resolve after the first lands. [11:31] maxb: Sure thing, will do. [11:33] can someone explain how does the authentication work in launchpad ? [11:33] Speaking of which, who would know what became of the mention of the potential of community PQM access? [11:38] asabil: it's all in canonical.launchpad.webapp.authentication [11:38] asabil: it's moderately complicated though [11:38] asabil: how good are you at zope? :-) [11:38] mwhudson, well I am trying to run launchpad, it works perfectly until I try to push code [11:38] mwhudson, not familiar with zope at all [11:38] oh [11:39] that's an entirely different thing [11:39] asabil: what user are you trying to push as? [11:39] it's the authentication in the xmlrpc service [11:39] a user I created [11:39] did you add an ssh key? [11:39] yes [11:40] where are the keys stored btw ? [11:40] in the database [11:40] let me check [11:40] asabil: can you ssh $USER@bazaar.launchpad.dev [11:40] ? [11:40] You'll probably need launchpad.dev:5022, actually. [11:40] let me check [11:40] yes, it's on 5022 [11:41] oh right yes [11:41] i have had that in ~/,ssh/config for so long... [11:42] mwhudson, the server is running but I get an auth failure [11:43] the console running 'make run_all' will have some messages [11:43] http://pastebin.com/d107872fc [11:44] that's the log from /var/log/launchpad/launchpad.log [11:44] er [11:44] that's strange [11:44] btw, it is a custom deployment [11:44] ah [11:44] Your Apache config is broken. [11:44] in other words I am fiddling with launchpad [11:44] I am using NGINX actually [11:45] yes, what wgrant says is certainly possible [11:45] so if I can just understand what's the problem [11:45] I can fix it [11:45] asabil: well, the ssh server will be talking to xmlrpc-private.launchpad.dev [11:45] private ? not xmlrpc.launchpad.dev ? [11:46] it looks like the virtual host config is busticated so the appserver can't tell which host the request was aimed at [11:46] asabil: yes, private [11:46] asabil: launchpad has two xml-rpc servers, a public one and a private [11:46] one [11:46] xmlrpc(-private) is served on a different port. [11:46] where is it supposed to be sent ? [11:46] localhost:8087 ? [11:46] or localhost:8085 ? [11:46] 8087 [11:47] both private and non private ? [11:47] i guess, look at configs/development/local-launchpad-apache or whatever the file is called [11:47] yes, I read it [11:47] I skipped the rewrite rules to get loggerhead working [11:47] but besides that the config is the same [11:48] Even the order? [11:49] http://pastebin.com/d2ab85831 [11:50] hmm, why is bazaar.launchpad.dev using a different local IP ? [11:50] i don't want to sound sceptical, but if it was exactly the same, i'd expect it to be working, and it's not working [11:50] 127.0.0.99 instead of 127.0.0.88 ? [11:50] i don't really remember [11:50] Does nginx not respect order? [11:50] 99 is only really used for private codebrowse. [11:50] Not xmlrpc. [11:51] mwhudson, I am not disagreeing with you, I also suspect it. I am just wondering why [11:51] wgrant, it does [11:51] asabil: xmlrpc-private.launchpad.dev listens on 80, not 443 [11:52] * mwhudson zzzs [11:52] mwhudson, in nginx I did setup both [11:53] asabil: Disable the .launchpad.dev directive and see what happens? [11:54] good morning launchpadders [11:54] Evening Ursinha. [11:54] wgrant: always this tz thing :) [11:54] wgrant: good evening then [11:55] wgrant, I just did doesn't change much [11:56] asabil: Check nginx and Launchpad logs to see which appserver is being forwarded to. [12:04] wgrant, http://pastebin.com/d76f839ed [12:04] that's what I see in wireshark [12:06] and that's what is proxified by nginx: http://pastebin.com/d3ee120b7 [12:07] asabil: To which IP address is the backend request? [12:07] 127.0.0.1 [12:07] Port, sorry. [12:08] nginx(80) ----> launchpad(8087) [12:09] That would appear to be correct. [12:09] anyway, thanks a lot for your help [12:09] I will continue investigating [12:10] You have not altered the Launchpad configuration at all? [12:10] may I know which host does launchpad expect ? [12:10] and where is it specified [12:10] wgrant, well I did as I said earlier [12:10] I want to understand how it works [12:11] I mean the Launchpad config, not the frontend HTTP server. [12:12] wgrant, yes I did [12:12] asabil: What have you changed? [12:13] wgrant, seems like I changed the auth endpoint line to xmlrpc instead of xmlrpc-private [12:13] I guess that's the issue [12:16] wgrant, thanks a lot [12:16] that was my mistake === mrevell is now known as mrevell-lunch [12:16] but I am still not able to login for some reason [12:17] oh that's normal, there is no shell [12:17] wgrant, thanks a lot, not it works [12:18] I just need to understand the various cron jobs to get everything setup now [12:32] gmb, re bug 519271 [12:32] Bug #519271: Ordering of latest bugs on the Malone frontpage appears to have become volatile [12:32] gmb, isn't that to be expected after dropping the 'id' from the query? [12:33] All bugs and bugtasks appear to have distinct creation dates, though. [12:44] oh, that bug would explain the spurious test fails I've seen [12:46] if two things are created in the same transaction, they have the same create time [12:46] I'm also intrigued as to how the extra order clause ended up so slow. [12:50] * bigjools adds MP to finally destroy Secure* publishing records [12:51] !!! [12:51] It hasn't bitrotted again? [12:51] nup, I kept it up to date, just didn't want to release it with the buildd changes last cycle [12:51] it's now 1 year and 1 week old :) [12:51] Excellent. [12:52] \m/ [12:52] and on that note, food. [13:00] jml: Well, no, because the creation dates in sampledata are distinct; the ordering in the tests should be sane. I need to investigate to be sure, though. [13:01] wgrant: Some investigation by jtv yesterday led us to the conclusion that postgres was vastly underestimating the number of rows returned by the query. Adding an index on (datecreated, id) increased the speed of the query by 4 OoM. [13:02] gmb, fair enough. perhaps the tests are using the factory? [13:02] gmb, zowie! [13:02] gmb: well I suspect that an underestimate (or buffer shortage) while sorting may be one of the causes. The other one is that you can skip the sorting altogether by adding this index. [13:02] gmb: Aha. Ew. [13:02] more of that please! [13:03] gmb: also, OOM was not in the GTF yet, so please join the GCP. http://xs4all.nl/~jtv/gtf/ [13:03] jml: D'oh. Looking at it now, that's exactly the problem. [13:03] jtv: SWD. [13:03] gmb: ? [13:03] jtv: Sure, will do. [13:03] * jtv adds another one... [13:04] and that's three-letter combination #9713 [13:04] (entry 23510) [13:04] gmb, fwiw, preliminary research indicates that transaction.commits in the factory are a big cause of slow tests. [13:05] jml: Wouldn't surprise me. [13:05] anyway, I want to eat something's flesh. === mrevell-lunch is now known as mrevell [13:57] Thankyou, bzr. http://pastebin.ubuntu.com/372497/ [13:57] I hope that is a bug. [14:00] up late there wgrant [14:02] bigjools: A little. [14:04] sinzui: Are you trying to start a riot in bug #519100? [14:04] Bug #519100: registry mails use male pronouns [14:05] wgrant: No. I am quietly stating that Launchpad made a decision that the registry must abide by [14:06] sinzui: Any such decision is probably ill thought out. [14:06] No other software in this world has such a policy. [14:07] indeed [14:10] sinzui: where is this writing guideline? [14:10] I've not seen it either. [14:11] wgrant: I think raging against how language evolves is very pointless. I think the masculine references to users software are a great method to alienate 51% of the world and prevents Launchpad from attracting a balanced group of users [14:11] sinzui: Then why is that bug Won't Fix? [14:12] bac, wgrant: They were among the documents that were supposed to move from the old wiki to the new wiki a few weeks ago. It was in the page where mpt gave example of improper and correct phrases [14:12] Ah, i think I've seen that one. [14:12] wgrant: 1) it is launchpad, not registry, 2) it is policy that I lost the battle on. [14:12] I cannot find it now, however. [14:13] Then the battle should be resumed and trivially won. [14:13] * bac searching [14:13] https://dev.launchpad.net/UserInterfaceWording [14:13] wgrant: Do you think we should pretend it was lost in the great London flood of 2008? [14:13] sinzui: in this specific case we could solve the problem and adhere to the policy by just removing the word 'himself' as the sentence is clear and corrrect without it. [14:14] That page doesn't say anything about pronouns (and I certainly never would have prescribed using "him" etc) [14:14] so maybe it's another page sinzui is referring to [14:15] mpt: There was a document on it because I wrote her and I was show the wiki page that said I had to use him [14:15] Using either is very probably a bug. [14:18] sinzui, would this have been something beuno wrote? I could ask him [14:20] beuno: would not have written it, I learned this rule in 2007 [14:20] i've searched the old wiki for 'pronoun', 'masculine', and 'male' and cannot find anything [14:21] bac: me too, and there is nothing. [14:21] bac: I am happy to declare this issue moot and will reopen the bug [14:21] sinzui: Thanks. [14:21] sinzui: i'd reopen and s/himself// as above [14:21] I will also stop wasting hours writing retarded sentences [14:22] * bac uh-oh, new can of worms... [14:22] Really, that message shouldn't use pronouns anyway [14:24] "The membership status of %s (%s) in the team %s (%s) was changed by the user himself from Approved to Deactivated." is just a long-winded way of saying "%s (%s) has left the %s team (%s)." [14:26] sinzui! you're alive! did you manage to get home uneventfully [14:27] beuno: I suppose I did, but I think there is a grand bogus saga waiting to be told about my last two weeks. It may arrive in warthogs in a week [14:28] It will not be as good as the beach trip to hell [14:28] sinzui, I'm looking forward to it [14:28] sinzui, also, I want to have a call with you at some point this week [14:28] I've been thinking about stuff [14:29] and there's one final thing I want to do before officially moving to U1 [14:29] and it involves you :) [14:29] beuno: fab [14:32] * mpt wonders why, about once a week, someone mistakenly links a branch to bug 1 [14:32] Bug #1: Microsoft has a majority market share

kfogel, are we on for a call on the hour? [15:01] jml: hey! [15:02] jml: been a while. Yes; short one, but let's do it. [15:02] jml: skype? [15:03] jml: did I miss my chance? [15:12] kfogel, you just dropped out. I'm reconnecting. === salgado is now known as salgado-lunch === matsubara is now known as matsubara-lunch === jamalta-afk is now known as jamalta === matsubara-lunch is now known as matsubara [16:39] jml: Thanks for the comment on the twisted-threading-bug-491870 merge proposal. Are you also implying that you think it's better to create the ThreadPool explicitly, or was it a FYI? === salgado-lunch is now known as salgado [17:54] allenap, I think FYI. I've made multiple comments, so I'm not sure which you mean. [17:55] jml: The most recent about adding an exception to importfascist. [17:55] allenap, oh yeah, that was an FYI. Better to shut up the warning _and_ file a bug on Twisted than do both. [17:55] anyway, I'm off. [17:55] jml: Thanks, cheerio. [17:56] np. g'night. === matsubara is now known as matsubara-afk === salgado is now known as salgado-afk [22:14] * thumper -> coffee [22:30] * mwhudson vanishes again === jamalta is now known as jamalta-afk